1 _____________________________________________________________________________
4 - Lay down the law on the different kinds of Sequence productions
7 - switch maximal to not-followed-by (~/~)
9 - should Union.add() be there?
10 - should Atom.top() be there?
12 - fix the location stuff, it's broken
13 - decent/better error messages
19 - RFC2822 (email message/headers)
22 ______________________________________________________________________________
25 - clean up the whole Walk situation
27 - cleaner solution to "maximal"?
30 - right now I can only lift the last child in a forest... begs
31 the question of what the right representation for Forests is
32 if we need to be able to do lift operations on it.
35 - "Regular Right Part" grammars (NP Chapman, etc)
36 - Attribute unification
38 - serialization of parse tables
39 - inference of rejections for literals
40 - "prefer whitespace higher up" (?)
41 - "ambiguity modulo dropped fragments"?
42 - can this be checked statically?
43 - eliminated statically?
45 ______________________________________________________________________________
48 - Implement a k-token peek buffer (for each state, see if it "dead
49 ends" during the next k Phases based solely on state -- ignoring
52 - Arrange for the SPPF corresponding to dropped subtrees to never be
53 generated (or merged, etc)
55 - Is there any way we can avoid creating a GSS.Node instance for
56 nodes which are transient in the sense that they have only one
59 - Implement "GLR syntactic predicates" -- the ability to do
60 arbitrary lookahead (ie "followed-by" and "not-followed-by" for
61 arbitrary patterns). This enables generalized longest-match and
62 lets us drop the Maximal hack.
64 - Re-read Rekers, particularly the stuff on optimal sharing
66 - Isolate the Element objects from Parse.Table/GSS so we can move
69 - consider allowing a Forest.Body to represent some other Tree whose
70 Body's should be [recursively] considered part of this Forest.
72 - perhaps not: right now we have a nice situation where
73 Forest.Ref instances become immutable once iterator()ed. This
74 also gives us a strong place to to culling with the certainty
75 that we won't throw out a Body which would later be salvaged
76 by some yet-to-be-added dependency.
78 - Figure out if there is a way to:
80 - allow unwrapping of children other than the very last one.
82 - fold repetitions into an array form in Forest, before
83 conversion to Tree. The major problem here is that multiple
84 tree-arrays are possible, all of different lengths. Worse,
85 even if they're all the same length, not all elements belong
86 in the same "possibility vector" as all others. You
87 essentially need a GSS to represent the array, which perhaps
88 is what the unfolded form was in the first place.
90 - Wikipedia grammar (needs to be both lexerless and boolean)
93 => Ordered Choice (";" operator)
95 - bring back in parse-table phase resolution of precedence (just
96 like associativity). This can be inferred from the use of ">"
97 when the rules are in one of these special forms:
108 where "_" is anything and "E" is the defining nonterminal.
109 Essentially what we're looking for is the situation where the
110 leftmost portion of one rule produces another rule, and the
111 rightmost portion of the latter produces the former.
113 I'm not 100% certain that this is as "strong" as the prefer/avoid
114 form (try to prove this, you probably can), but it's "what people
115 intend" most of the time.
117 - implement Johnstone's algorithm for "reduced, resolved LR
118 tables" to eliminate superfluous reductions on